iiinova — where your digital self keeps evolving

Reposted by Duarte David

Tom Forsyth @tomforsyth.bsky.social · 12/10/2026

DASHR - Dynamically Animated Skinned Heightfield Rendering A new rendering algo that lets you merge a whole bunch of ShaderToy and raytracer-style techniques right into your existing bone-animated triangle mesh rasterisation pipeline. New longer technical details video: youtu.be/SA2cXydaq1A

youtu.be

DASHR - Dynamically Animated Skinned Heightfield Rendering - technical details.

YouTube video by Tom Forsyth

211731

Reposted by Duarte David

Alan Wolfe @demofox.bsky.social · 01/10/2026

df_game_data v1.0.0 released! github.com/Atrix256/df_... An editor and data compiler for read only game data in C++. Describe schemas using C style stucts, edit data in the editor, compile data to make a .bin file and a generated .h, use the .h to load the .bin in your code, make your game.

A small C++ program showing how to use the bin file and generated header file that comes out of df_game_data when you compile data.

df_game_data editor showing a character table being edited. it has 3 records: curly, larry and moe. The larry record is being edited.

An example .def schema file. It uses C/C++ style syntax for describing data structures. Specify the table entry type using #root.

26515

Reposted by Duarte David

Sophie Huiberts @sophie.huiberts.me · 28/09/2026

The most important principle when teaching LP is to not cover any theoretical nonsense

sophie.huiberts.me

Principles for Teaching Linear Programming

4223

Reposted by Duarte David

Tom Forsyth @tomforsyth.bsky.social · 28/09/2026

DASHR - Dynamically Animated Skinned Heightfield Rendering Short video for now - I might do a longer one later? youtu.be/-Su9YrcazRk But there's ALL the details in the paper, and you can always Use The Source, Luke for the really deep info. github.com/tomforsyth10...

youtu.be

DASHR - Dynamically Animated Skinned Heightfield Rendering - six minutes version

YouTube video by Tom Forsyth

1423377

Reposted by Duarte David

Ash @fishyash.com · 21/09/2026

Friendly reminder that Duolingo was initially concieved to trick people into translating articles for free, which they later used as training data for MTL algorithms. Duolingo's primary purpose has NEVER been about teaching you a language, even from the start! slate.com/technology/2...

If it doesn’t charge users and, so far, doesn’t serve them any ads, how does Duolingo make money? It tricks you into working as an unpaid translation service. At the end of some lessons, Duolingo will ask you if you want to practice by translating a real-world document. Which will turn out to be, say, a BuzzFeed article. By melding together enough stabs at this task from high-level Duolingo users, the app can render a surprisingly accurate translation. Which is worth good money. So far, Duolingo has contracts with BuzzFeed and CNN to translate stories from English into languages such as Spanish, French, and Portuguese. This service is already earning Duolingo hundreds of thousands of dollars a year, but von Ahn predicts he will be adding a raft of new clients soon. “The language translations market is huge,” he says. “It’s a $30 billion per year market.”

2227131667

Reposted by Duarte David

doctober @docseuss.bsky.social · 29/08/2026

ken penders may be the only person to perfectly desecrate martin neimoller's memory

46811

Reposted by Duarte David

Robin Green @fatlimey.bsky.social · 14/08/2026

Yes, that too. mandyxia.com/research/ass...

$A row of 5 diagrams showing different techniques for rendering hair reflection, only a few take into account diffraction.$

031

Reposted by Duarte David

Robin Green @fatlimey.bsky.social · 14/08/2026

Quite a lot of the scattering in fur volumetric shadows is from diffraction around the hair strands. A lot more complicated than sums of random paths inside a medium, also frequency dependent.

171

Reposted by Duarte David

Maximilian Tarpini @maxtarpini.bsky.social · 13/08/2026

All currently available SSS models ignore internal boundary condition (aka TIR) .. but trapping rays past the critical angle (keeping'em bouncing) forces longer path lengths .. for richer, more saturated, physically accurate translucency. Same color and mfp. With / without TIR

2449

Reposted by Duarte David

Ipotrick @ipotrick.bsky.social · 08/08/2026

Pioneer Pretraced rayguiding! * Shoot low res rays * Spatially spread bright ray hits * Use bright rays to guide real traced rays Drastically improved directional GI quality, low cost, around 0.2-0.3ms extra Inspired by Restir, but this has 0 additional temporal lag!

1356

Reposted by Duarte David

Joshua Bowman @thalesdisciple.bsky.social · 08/08/2026

Working on an A.I. policy for my calculus classes. Decided to be upfront about the current situation with LLMs and to give a heartfelt explanation for why the course still matters.

Al policy

Context

Artificial Intelligence, or "A.l.," Is a broad term used to refer to several different kinds of technologies. The kind most relevant to this class are Large Language Models (LLMs), which use mathematics to analyze large amounts of human-produced text and then generate new text that simulates what a human might write in response to prompts or questions. Some examples of LLMs are OpenAl's ChatGPT, Anthropic's Claude, and Google's Gemini.

Current LLMs can provide a solution to any calculus problem I give you. So why should we work on them, you might ask? Here's the thing: for the most part, the problems were already solved anyway. Calculus as a subject is nearly 400 years old. It has reached into virtually every modern scientific and economic field. And while I would like to believe that I have crafted a few new problems with some small amount of originality tailored to this class, for the most part I have curated a collection of exercises drawn from my years of teaching calculus. Those exercises come from textbooks, workshops, friends, and other fields of study, as well as my own invention.
And they nearly all have known solutions, often with fairly standard approaches.
Before A.I. became widely available, you could find those solutions in another textbook, or from a friend's notes, or through a repository on the internet.

The point of this class is not necessarily to solve problems that have never been solved before. It is for you to develop habits of thinking that are used by mathematicians, scientists, and policymakers-to develop a sort of mental toolbox that can be applied to new problems. If along the way you happen to find a new way of solving one of the class exercises, that's great! I am constantly learning new things from my students. But you will be bereft of discovery if you rely on a LLM for solutions to the calculus problems before you. Here I mean the discovery of having your own realizations, crystallizing your own internal understanding, claiming the knowledge in a way that cannot be taken away from you.
That ownership is what I want for you, and that is what the course is built around.

441923545

Reposted by Duarte David

Claire Blackshaw @evilkimau.bsky.social · 06/08/2026

Slowing skimming Siggraph, this one stood out. Not the main point but this approach solves the "ghostly" vibe by having coarse rep with freq wave on top. Which also gives you continuous LOD without holes. Not VR frametimes yet, but interesting ideas. arcanous98.github.io/projectPages...

arcanous98.github.io

Gabor Fields: Orientation-Selective Level-of-Detail for Volume Rendering

Gabor Fields — orientation-selective level-of-detail for volumetric rendering with Gabor kernel mixtures.

021

Reposted by Duarte David

Xiana Carrera @xianacarrera.bsky.social · 04/08/2026

The talk I presented at #SIGGRAPH2026 is now available on YouTube, for anyone who couldn’t attend or would like to revisit it :) youtu.be/OFJYBUArVPc?...

youtu.be

Dual Contouring of Signed Distance Data - SIGGRAPH 2026 [Technical paper presentation]

YouTube video by Xiana Carrera

27219

Reposted by Duarte David

ceruleite.bsky.social @ceruleite.bsky.social · 03/08/2026

Slides for SIGGRAPH 2026 Advances in Real-Time Rendering in Games course's talk on SMOLDER — REAL-TIME VOLUMETRIC EFFECT RENDERING IN GLACIER AND 007 FIRST LIGHT by @lexdev.net from IO Interactive are now online - enjoy!! advances.realtimerendering.com/s2026/index....

05419

Reposted by Duarte David

Mischief @mischief.games · 30/07/2026

Welcome To The New Apocalypse

411565363

Reposted by Duarte David

Daniel Holden @theorangeduck.bsky.social · 30/07/2026

One of my most requested articles is something on inverse kinematics and foot sliding. I finally found the time to write up the little recipes that have served me well through the years, including an offline method I used recently that works surprisingly nicely. theorangeduck.com/page/inverse...

03711

Reposted by Duarte David

Nicole Feng @nicole--f.bsky.social · 28/07/2026

At #SIGGRAPH2026, I presented "points as tori" (PAT) for fast pointwise evaluation of signed distance for point clouds. Some code (Python w/ C++ extensions) and demo up: github.com/nzfeng/point... Contributions from people/LLMs welcome. Explainer thread below for the project: 🧵

411925

Reposted by Duarte David

Yining Karl Li @yiningkarlli.bsky.social · 25/07/2026

Also, in retrospect, the Gradient Domain Path Tracing paper by Kettunen et al. 2015 was also an extraordinarily high impact paper. I didn’t recognize it at the time, but it opened the door to all of the differentiable and invertible approaches that now underpin the neural rendering subfield.

041

Reposted by Duarte David

Bruno Levy @brunolevy01.bsky.social · 23/07/2026

Oops, I completely forgot to print my poster for the Q&A session after the SIGGRAPH presentation... Let us try to fix this...

2235

Reposted by Duarte David

Dominic Tarason @dominictarason.com · 23/07/2026

GOD DAMMIT. If publishers were worth a damn these days, they'd have funded whatever they wanted to make next. They were a small studio but with an environmental art and character animation team better than most AAA outfits. Crisol's world is absurdly detailed.

927055

Reposted by Duarte David

Nicole Feng @nicole--f.bsky.social · 19/07/2026

Tomorrow at #SIGGRAPH2026 (Monday) @ 3:45-5:35pm, come to the papers session "Creating & Iso-Surfacing SDFs" in Room 408B. A paper I'm looking forward to is "Subgrid Marching Tetrahedra" by Hossein Baktash (yes, I'm biased as his labmate, but it's a great idea) [1/n]

194

Duarte David @duartedavid.bsky.social · 22/07/2026

Took an evening to turn my bidirectional pathtracer into a transient renderer :)

150

Reposted by Duarte David

Tzu-Mao Li @tzumaoli.bsky.social · 20/07/2026

Hopefully one last SIGGRAPH advertisement. I, myself, yours truly, will present on Haolin's behalf a data-driven method for joint adaptive sampling and quadrature for integration. Wed R408A. In the end of the talk I'll argue why everyone is doing the wrong thing. Paper: suikasibyl.github.io/nqr

0156

Reposted by Duarte David

David Levin @diwlevin.bsky.social · 20/07/2026

Fun poster at #SIGGRAPH2026 with incoming PhD student Yixin Lok. Sadly the paper is there but Yixin isn't due to visa issues.

1133

Reposted by Duarte David

Yining Karl Li @yiningkarlli.bsky.social · 20/07/2026

Disney Animation has a bunch of talks at #SIGGRAPH2026; each day I'll post a thread about our talks for the day! In total my coworkers and I are presenting a total of 10 talks. Here are the talks my coworkers are presenting Sunday/Monday: (1/10) disneyanimation.com/events/?draw...

disneyanimation.com

Walt Disney Animation Studios - Events

The official site of Walt Disney Animation Studios.

1203

Reposted by Duarte David

Ahmed Mahmoud @ahdhn.bsky.social · 18/07/2026

Next week at #SIGGRAPH2026, we'll present "Fast Sparse Matrix Permutation for Mesh-Based Direct Solvers," with Behrooz Zarebavani, Ana Dodik, Changcheng Yuan, Serban Porumbescu, John Owens, Maryam Mehri Dehnavi, and @justinmsolomon.bsky.social . Talk: Wed., July 22, 3:55 PM, Room 403-B (1/6)

1116

Duarte David @duartedavid.bsky.social · 18/07/2026

Imagine that the only good product your company has is providing servers for an open source app, and feeling like you know enough to talk about anything else.

010

Reposted by Duarte David

Tzu-Mao Li @tzumaoli.bsky.social · 17/07/2026

For people coming to SIGGRAPH next week: Xuan will present this paper on Thursday 9:50am in the "Prevent Collision" session in Room 403A. Come chat with us!

092

Reposted by Duarte David

pltc @pltc.bsky.social · 16/07/2026

Packing my bag for #SIGGRAPH2026! @brunolevy01.bsky.social and I will be presenting our course on Optimal Transport for Fluid Simulation on Sunday July 19, 9am-12:15pm, Room 408B! Come say hi if you're around!

0254

Reposted by Duarte David

ceruleite.bsky.social @ceruleite.bsky.social · 15/07/2026

2026 Advances in Real-Time Rendering in Games course line up is here - check it out! advances.realtimerendering.com

16127

Reposted by Duarte David

Ahmed Mahmoud @ahdhn.bsky.social · 14/07/2026

Next week at #SIGGRAPH 2026, we'll present "Locality-Aware Automatic Differentiation on the GPU for Mesh-Based Computations," with Rahul Goel, Jonathan Ragan-Kelley, and @justinmsolomon.bsky.social . Talk: Mon., July 20, 2:50 PM, Room 403-B Paper: dl.acm.org/doi/10.1145/... More below (1/7):

22111

Reposted by Duarte David

Claire Blackshaw @evilkimau.bsky.social · 13/07/2026

Before you go okay granny. AMD RDNA 3 L1 cache is often 32kb with L2 being in 4-64mb range depending. Q2 has 1mb total mem per tile with Q3 era hmds bumping it up to 2mb I'm not saying we can't have nice things but physics still exsists!

142

Reposted by Duarte David

Ben Beilharz @engramm.ing · 06/07/2026

Pretty cool new sim work by TUM: www.youtube.com/watch?v=-1Tx... Paper: ge.in.tum.de/download/ST-...

youtube.com

SIGGRAPH'26: Spatiotemporal FLIP for Fast Free-Surface and Two-Phase Sims With Very Large Time Steps

YouTube video by Nils Thuerey

084

Reposted by Duarte David

Amir Vaxman @amirvaxman-dgp.bsky.social · 03/07/2026

We won best paper at SGP! Special congratulations to Hongyi, who unfortunately could not make it,

0141

Reposted by Duarte David

Yining Karl Li @yiningkarlli.bsky.social · 02/07/2026

In this thread: three cool rendering papers that DisneyResearch|Studios is publishing at EGSR this week. The authors of these papers are peers that I get the privilege of working closely with frequently, and these are projects I've been excited to watch develop! (1/17)

13412

Reposted by Duarte David

UNITED24 @u24.gov.ua · 02/07/2026

The number of victims has risen to at least 17💔

15021

Reposted by Duarte David

Paul "Paulie" Hughes @pauliehughes.bsky.social · 01/07/2026

Dang! This is going to be great! Add to basket.

082

Reposted by Duarte David

Cem Yuksel @cemyuksel.com · 01/07/2026

Our #HPG2026 paper "Memory-Efficient Bounding Volume Hierarchies with Merged Nodes for Hardware Ray Tracing" introduces new ways of constructing BVHs that consistently outperform state-of-the-art BVHs for GPU ray tracing. Project page: graphics.cs.utah.edu/research/pro... @ebrunvand.bsky.social

25415

Reposted by Duarte David

kyndinfo @kyndinfo.bsky.social · 28/06/2026

Smoothing スムージング [Math short write-up] www.kynd.info/writing/smoo...

0405

Reposted by Duarte David

Xiana Carrera @xianacarrera.bsky.social · 18/06/2026

1/6 Excited to share our #SIGGRAPH2026 paper, Dual Contouring of Signed Distance Data! We propose a new method for reconstructing sharp meshes from discrete samples of signed distance functions (SDFs) that does not require gradients. Project page: gatc.cs.columbia.edu/projects/dua...

410626

Reposted by Duarte David

Nicole Feng @nicole--f.bsky.social · 15/06/2026

I recently completed my PhD thesis, where I developed algorithms for making geometric inferences from imperfect data (i.e. incomplete and inaccurate observations/representations of shapes, like broken meshes or point clouds). A thread 🧵:

25516

Reposted by Duarte David

Megan Fox @glassbottommeg.bsky.social · 07/06/2026

I don't know about you, but when I'm trying to find a tutorial for something I ask a friend "hey what's a good tutorial" and they say "oh try this one it was great". Or I read the comments. Which are vicious and hound authors for even slight errors. Or I use a course/book from a reputable source.

4111

Reposted by Duarte David

Joe Wintergreen @joewintergreen.com · 28/05/2026

THEY FUCKING CHANGED IT. KINGS AND QUEENS

814424

Reposted by Duarte David

Tzu-Mao Li @tzumaoli.bsky.social · 25/05/2026

github.com/txstc55/yasps Check out Xuan's work on a new programming language for writing modular and high-performance physical simulation code from energy definition. The key is two differentiable relational operators (join/union) that model the sparsity structure between the attributes and energy.

0111

Reposted by Duarte David

Jacco Bikker 🎦 @jbikker.bsky.social · 22/05/2026

markjgillespie.com/Research/Sub... Via: www.realtimerendering.com/kesen/sig202... Even if you're not interested, read it for the incredible visual design of that page. I'm talking about the first character, how is that called?

markjgillespie.com

Subgrid Marching Tetrahedra

2142

Reposted by Duarte David

Christopher Batty @topher-batty.mastodon.acm.org.ap.brid.gy · 20/05/2026

We have a fun new paper to appear at SIGGRAPH 2026, led by Columbia PhD student Xiana Carrera, that extends dual contouring to work nicely with [pure] signed distance samples. It avoids classic DC's additional requirement of providing accurate Hermite / […] [Original post on mastodon.acm.org]

A 3D model of a fan disk, reconstructed from signed distance data using (left) classic dual contouring with finite difference estimates of surface normals/gradients, which shows bumps and unevenness near expected sharp curved features, and (right) our improved method, which produces much cleaner and sharper reconstruction of features.

14011